Introduction

Radiofrequency ablation (RFA) was first used to treat back pain in 1975 by Shealy. The procedure involves running high-frequency electrical current through an insulated needle, resulting in thermal energy at the tip of the needle. The needle is positioned such that the heat from the tip of the needle produces a small lesion within the target nerve, disrupting pain signals from the nerve. Ablation most often occurs through 21- or 22-gauge needles for 1 to 3 minutes at 75 ^° C to 90 ^° C. Over time, modifications and variations to the traditional RFA technique have been developed. Cooled RFA, first described in 1996, uses water circulation around the electrode tip to create a larger, spherical ablative area of approximately 60 ^° C around the needle tip, allowing the proceduralist to compensate for anatomic variability and probe angulation while still ablating the target structure. ^, Pulsed radiofrequency (PRF) is a newer technique that uses intermittent administration of current, resulting in temperatures below 42 ^° C, thus preventing neuronal damage. Though the mechanism is not fully understood, PRF may have neuromodulatory effects at the targeted site.

Traditional RFA has been used to treat multiple sources of chronic back and joint pain such as facet joint pain, sacroiliac joint pain, degenerative disc disease, and cervical radicular pain. ^, Although initially studied for knee osteoarthritis and sacroiliac joint pain, new evidence indicates that cooled RFA is as efficacious as traditional RFA for thoracic and lumbar facet joint pain. ^, While large randomized clinical trials are lacking, RFA has been shown to lead to a short-term improvement in pain in patients with chronic low back pain. Additionally, RFA is more effective for short- and long-term pain than steroid injections. Although typically well tolerated, RFA is not the solution to all pain syndromes and requires careful patient selection. In this chapter, we will describe a number of preprocedural, intraprocedural, and postprocedural considerations when performing RFA.

Preprocedural

As with any procedure, the first part of a patient evaluation is a thorough history and physical to narrow down the source and nature of the pain. However, a history and physical are often not sufficient to diagnose the etiology of chronic low back pain, and diagnostic blocks should be used for patient selection. Once a patient has been deemed an appropriate candidate for RFA, a variety of patient-related factors should be considered that may affect the procedure.

Patients on antiplatelet or anticoagulation medications present a challenge due to the risk of bleeding-related complications, such as epidural hematoma. However, this risk must be balanced with the risk of thrombotic cardiovascular or cerebrovascular events in the absence of anticoagulation. Aspirin and nonsteroidal antiinflammatory drugs (NSAIDs) appear to be safe for spine interventions and should not be discontinued for spine or sacroiliac joint procedures. Due to limited studies on the safety of P2Y12 inhibitors (clopidogrel, ticagrelor, etc.) and novel oral anticoagulants (NOACs), multiple international societies collaborated to release a comprehensive set of guidelines for anticoagulation and interventional spine procedures. In general, for cervical spine RFA, P2Y12 inhibitors should be held for 5 to 7 days prior to the procedure while NOACs should typically be held for five half-lives prior to the procedure. Thoracic and lumbar spine RFAs are considered lower risk; as such, careful patient-specific risk assessments are warranted when considering continuing or discontinuing anticoagulation prior to the procedure per ASRA guidelines. Finally, radiofrequency neurotomy is contraindicated in patients with significant bleeding diathesis or systemic or local infections.

Special care should be taken in patients with preexisting implanted devices, such as deep brain stimulators (DBSs), spinal cord stimulators (SCSs), pacemakers (PPMs), and defibrillators (AICDs). In general, bipolar radiofrequency (RF) is recommended in patients with implanted devices, as this limits electrical current through the patient’s body and thus the device. Thermal damage can occur from heat transfer from the RF instrument to the implantable device, though the risk is lower the farther away the site of ablation. For example, a complete vegetative state has been reported from thermal heat transfer to a deep brain stimulator following RFA for maxillary osteonecrosis, but patients with DBS have undergone lumbar RFAs with no adverse events. Most common spinal cord stimulator manufacturers recommend against RFAs in patients with these devices, though case reports and retrospective studies recommend reprogramming the device for an output of 0 Volts, turning it off prior to the procedure, and avoiding contact with the leads to ensure safety. ^, Successful and safe RFAs has been performed in patients with PPMs and AICDs, though a multidisciplinary discussion between the patient, cardiologist, device manufacturer, and pain physician is recommended to determine device settings prior to the procedure (for example, an asynchronous pacing mode for the PPM). For these patients, an external pacing source should be readily available as well. Finally, any patient with an implantable device should receive minimal to no sedation to allow for prompt patient communication about any cardiovascular or neurologic symptoms.

Intraprocedural

One of the foremost considerations while performing RFA is whether or not sedation is warranted; and if so, what level of sedation is most appropriate. Whether or not sedation is used affects not only the patient’s experience of the procedure, but factors into its safety. Currently, the American Society of Interventional Pain Physicians (ASIPP) recommends that mild-to-moderate sedation be administered for all pain management procedures, as these levels of sedation are associated with increased patient comfort while also allowing for patient feedback on acute changes. ^, Sedation may be especially beneficial for patients with anxiety-related disorders and chronic pain conditions. That said, mild-to-moderate sedation was found not to impact postprocedure pain scores. In addition, sedation with benzodiazepines such as diazepam has not been associated with increased patient satisfaction after cervical or lumbar interventional pain procedures, with more than 90 percent of patients being either happy with receiving sedation or satisfied with the decision not to receive sedation. It is also well-recognized that patients with airway obstruction physiology, increased body habitus, and/or neurocognitive impairment can be high risk during even mild-to-moderate sedation. Therefore, it is crucial to evaluate the sedation’s intended benefits during the procedure itself (anxiolysis, positional comfort, and relative immobility) in the context of the patient’s risk factors, as opposed to focusing on the patient’s satisfaction with the experience postprocedure. Finally, it should be emphasized that, in the event that moderate sedation is to be utilized, a provider with experience in rescue from at least deep sedation ought to be present, in accordance with practice guidelines from the American Society of Anesthesiologists.

Other notable concerns include pain and injury during RFA. While continuous RFA has been shown to outperform pulsed RFA for improvement in pain symptoms and functionality after treatment of facet join pathology, it has also been associated with a greater incidence of localized pain at the lesion site, dysesthesias, and neuritic pain. This appears not to be the case for lumbar radicular pathology. Vascular penetration and injury are of concern during the procedure, given an increased likelihood of epidural hematoma as the dreaded result. The incidence of vascular injury has been estimated at 6.1%, particularly for procedures involving the medial branches. In situations where vascular penetration is suspected, the aspiration test is 34.1% sensitive for detecting the event, while performing spot radiographs under administration of contrast is 59.1% sensitive. Other potential complications of RFA include superficial burns and postprocedure complications that are discussed in the following section; however their global incidence is estimated to be 1%. Cooled RFA, a newer modality, for which there is more literature on the treatment of osteoarthritis-related knee pain, has been gaining traction for the treatment of back pain as well. Its use specifically for sacroiliac joint pain has been associated with only transient complications such as hip soreness and numbness.

Postprocedural

A number of outcome- and safety-related factors are important to address so as to ensure proper success and follow-up after RFA. In the event that sedation is used during the procedure, the patients should be monitored on site for residual sedation-related effects. Adverse events, though relatively rare (1% or less), should be anticipated. Patients may experience subsiding pain, transient lower limb weakness, nonpainful paresthesias, and even exacerbation of their preprocedural pain. These events are thought to be at least partly mediated by proinflammatory cytokine activation during tissue ablation and may be minimized by the local injection of corticosteroid at the RFA site prior to cannula removal. Finally, while infection following RFA is a theoretical concern, its reported incidence has been essentially zero.

For monitoring treatment effect, success has been commonly defined as at least a 50% improvement in pain symptoms by the 0–10 numerical rating scale (NRS) 3 months or more after the procedure. ^{, ,} This is a useful benchmark to discuss with the patient before the procedure and at follow-up because it sets appropriate expectations for treatment benefit and aids with structuring the follow-up schedule. Alternatively, given that the minimal clinically significant difference in NRS among patients with low back pain has been estimated to be two points (for example, going from a 9 to a 7, or an 8 to a 6), the two-point threshold can be used as an alternative benchmark for patients with chronic and/or multifactorial pain. Other instruments with published estimates of clinically significant difference for pain and functional status in the low back pain population include the Visual Analogue Scale (0–100), the Roland Disability Questionnaire (0–24), the Oswestry Disability Index (0–100), the Quebec Back Pain Disability Questionnaire (0–100), and the Euroqol 5D Health Questionnaire (0–1). While some of these instruments have been more commonly used in clinical trials and other research settings, they may be of benefit as adjuncts to the NRS for pain assessment and to capture additional dimensions of RFA’s treatment effect, including changes in functional status and global quality of life.